Gross Tumor Volume Centroid Research Articles

Volumetric MRI with sparse sampling for MR-guided 3D motion tracking via sparse prior-augmented implicit neural representation learning.

Volumetric reconstruction of magnetic resonance imaging (MRI) from sparse samples is desirable for 3D motion tracking and promises to improve magnetic resonance (MR)-guided radiation treatment precision. Data-driven sparse MRI reconstruction, however, requires large-scale training datasets for prior learning, which is time-consuming and challenging to acquire in clinical settings. To investigate volumetric reconstruction of MRI from sparse samples of two orthogonal slices aided by sparse priors of two static 3D MRI through implicit neural representation (NeRP) learning, in support of 3D motion tracking during MR-guided radiotherapy. A multi-layer perceptron network was trained to parameterize the NeRP model of a patient-specific MRI dataset, where the network takes 4D data coordinates of voxel locations and motion states as inputs and outputs corresponding voxel intensities. By first training the network to learn the NeRP of two static 3D MRI with different breathing motion states, prior information of patient breathing motion was embedded into network weights through optimization. The prior information was then augmented from two motion states to 31 motion states by querying the optimized network at interpolated and extrapolated motion state coordinates. Starting from the prior-augmented NeRP model as an initialization point, we further trained the network to fit sparse samples of two orthogonal MRI slices and the final volumetric reconstruction was obtained by querying the trained network at 3D spatial locations. We evaluated the proposed method using 5-min volumetric MRI time series with 340ms temporal resolution for seven abdominal patients with hepatocellular carcinoma, acquired using golden-angle radial MRI sequence and reconstructed through retrospective sorting. Two volumetric MRI with inhale and exhale states respectively were selected from the first 30s of the time series for prior embedding and augmentation. The remaining 4.5-min time series was used for volumetric reconstruction evaluation, where we retrospectively subsampled each MRI to two orthogonal slices and compared model-reconstructed images to ground truth images in terms of image quality and the capability of supporting 3D target motion tracking. Across the seven patients evaluated, the peak signal-to-noise-ratio between model-reconstructed and ground truth MR images was 38.02±2.60dB and the structure similarity index measure was 0.98±0.01. Throughout the 4.5-min time period, gross tumor volume (GTV) motion estimated by deforming a reference state MRI to model-reconstructed and ground truth MRI showed good consistency. The 95-percentile Hausdorff distance between GTV contours was 2.41±0.77mm, which is less than the voxel dimension. The mean GTV centroid position difference between ground truth and model estimation was less than 1mm in all three orthogonal directions. A prior-augmented NeRP model has been developed to reconstruct volumetric MRI from sparse samples of orthogonal cine slices. Only one exhale and one inhale 3D MRI were needed to train the model to learn prior information of patient breathing motion for sparse image reconstruction. The proposed model has the potential of supporting 3D motion tracking during MR-guided radiotherapy for improved treatment precision and promises a major simplification of the workflow by eliminating the need for large-scale training datasets.

Real Time Volumetric MRI for MR-Guided 3D Motion Tracking via Sparse Prior-Augmented Neural Representation Learning

To reconstruct volumetric MRI from orthogonal cine acquisition aided by sparse priors of 2 static 3D MRI through implicit neural representation (NeRP) learning, with the goal of eliminating large-scale training datasets for data-driven sparse MRI reconstruction and supporting clinical workflow of real time 3D motion tracking during MR-guided radiotherapy. A multi-layer perceptron network was trained to learn the NeRP of a patient-specific MRI dataset, where the network takes 4D data coordinates of voxel locations and motion states as inputs and outputs corresponding voxel intensities. By first learning the NeRP of 2 static 3D MRI with different breathing motion states, prior knowledge of patient breathing motion was embedded into network weights through optimization. The prior knowledge was then augmented from 2 to 31 motion states by querying the optimized network at interpolated/extrapolated motion state coordinates. Starting from the prior-augmented network as an initialization point, the network was further trained using sparse samples of 2 orthogonal cine slices. The final volumetric reconstruction was obtained by querying the trained network at desired 3D spatial locations. We evaluated the proposed method using 5-minute volumetric MRI time series with 340 ms temporal resolution collected from 7 liver carcinoma patients. The time series was acquired using golden-angle radial MRI sequence and reconstructed through retrospective sorting. Two MRI with inhale and exhale states respectively were selected from the first 30 sec of the time series for prior embedding and augmentation. The remaining 4.5-min time series was used for volumetric reconstruction evaluation, where we retrospectively subsampled each MRI to 2 orthogonal slices and compared network-reconstructed images to ground truth images in terms of image quality and the capability of supporting 3D target motion tracking. Across the 7 patients evaluated, the peak signal to noise ratio between model reconstruction and ground truth was 54.66 ± 6.16 dB and the structural similarity index measure was 0.99 ± 0.01. Gross tumor volume (GTV) contours estimated by deforming a reference state MRI to model-reconstructed and ground truth MRI showed good consistency. The 95-percentile Hausdorff distance between GTV contours was 1.89 ± 1.13 mm, which is less than the voxel dimension. The mean GTV centroid position difference between ground truth and model estimation was less than 1 mm in all 3 orthogonal directions. Volumetric MRI from orthogonal cine acquisition with sparse priors is feasible by modeling prior knowledge through implicit neural representation learning. The model-reconstructed images showed sufficient accuracy in supporting 3D motion tracking of abdominal targets. By eliminating the need for large scale training datasets, the method promises to enable clinical implementation of 3D motion tracking for precision radiation therapy.

Deep learning model fusion improves lung tumor segmentation accuracy across variable training-to-test dataset ratios.

This study aimed to investigate the robustness of a deep learning (DL) fusion model for low training-to-test ratio (TTR) datasets in the segmentation of gross tumor volumes (GTVs) in three-dimensional planning computed tomography (CT) images for lung cancer stereotactic body radiotherapy (SBRT). A total of 192 patients with lung cancer (solid tumor, 118; part-solid tumor, 53; ground-glass opacity, 21) who underwent SBRT were included in this study. Regions of interest in the GTVs were cropped based on GTV centroids from planning CT images. Three DL models, 3D U-Net, V-Net, and dense V-Net, were trained to segment the GTV regions. Nine fusion models were constructed with logical AND, logical OR, and voting of the two or three outputs of the three DL models. TTR was defined as the ratio of the number of cases in a training dataset to that in a test dataset. The Dice similarity coefficients (DSCs) and Hausdorff distance (HD) of the 12 models were assessed with TTRs of 1.00 (training data: validation data: test data = 40:20:40), 0.791 (35:20:45), 0.531 (31:10:59), 0.291 (20:10:70), and 0.116 (10:5:85). The voting fusion model achieved the highest DSCs of 0.829 to 0.798 for all TTRs among the 12 models, whereas the other models showed DSCs of 0.818 to 0.804 for a TTR of 1.00 and 0.788 to 0.742 for a TTR of 0.116, and an HD of 5.40 ± 3.00 to 6.07 ± 3.26mm better than any single DL models. The findings suggest that the proposed voting fusion model is a robust approach for low TTR datasets in segmenting GTVs in planning CT images of lung cancer SBRT.

Lung Tumor Motion and its Impact on Deep Learning Prediction of Local Recurrence

<h3>Purpose/Objective(s)</h3> In order for deep learning (DL) models to successfully predict local recurrence of lung tumors from image-based features, uncertainties of the model must be quantified. One potential source of uncertainty for the DL model is lung tumor motion. We aimed to evaluate the association of motion magnitude with the prediction accuracy of the DL model. We hypothesized that knowledge of motion will help quantify the uncertainties of prediction and guide the use of individualized risk predictions for lung radiotherapy. <h3>Materials/Methods</h3> Physician delineated gross tumor volume (GTV) of 19 patients were propagated to their corresponding 10 breathing phases of the 4DCT images using a commercial treatment planning system. A programming environment code was used to extract the contours and determine the centroid and volume bounded by the contours. Motion was determined from the displacement of the GTV centroid. The Euclidean distance was calculated and used as an aggregate measure of tumor motion magnitude. The 4DCT, free-breathing CT (FBCT) and maximum-intensity projection (MIP) images of the 18 patients were used as input to a pre-trained DL model. The 8-layer DL model was previously trained to predict local recurrence using FBCT images of 933 lung patients. Correlation between motion magnitude and prediction accuracy was calculated for 4DCT images in each respiratory phase, FBCT and MIP images. <h3>Results</h3> Tumor movement with a mean exertion of 1 cm [IQR: 0.72-1.53] was observed from the cohort with 47.4 % of patients subjected to abdominal compression, and 57.9% of patients had peripheral lung tumors. The standard deviation of DL scores obtained from each 4DCT image correlated with the magnitude of tumor motion (R² =0.45, p=0.005). Higher uncertainty in DL scores was attributed to the larger tumor motion. The average DL score obtained from the 10 phases of the 4DCT images of each patient was (i) highly associated with the DL scores obtained from the FBCT images (r = 0.99; p<0.0001), and (ii) closest to the DL score obtained with the end-inspiration images. In comparing the prediction performance across image acquisition types, MIP images yields a higher DL score compared to FBCT (<i>p</i><0.0001) or 4DCT images (<i>p</i> =<0.0001). <h3>Conclusion</h3> The impact of tumor motion on the range of DL scores for prediction of local recurrence was reported for the first time. Standard deviations of the DL scores increased with motion magnitude. However, when DL score was measured and averaged across the 10 phases of the 4DCT images, DL score correlated well with those obtained from the FBCT images. This preliminary finding suggests that despite the increased variation with larger motion, the DL score, when averaged across the sinusoidal breathing trajectory, correlates with the DL score obtained from the FBCT images that were used for treatment planning. However, caution should be exercised when using MIP images for local recurrence prediction since a higher DL score than the FBCT images used for planning will be observed.

Real Time Volumetric MRI Using a Physics-Aware Deep Neural Network for MR-Guided Radiation Therapy

To investigate the feasibility of real time volumetric MRI using a physics-aware deep neural network, in support of MRI-based 3D motion tracking for MR-guided adaptive radiotherapy.A deep neural network was proposed to generate volumetric MRI from orthogonal Cine MRI, which can be acquired on MR-Linac systems real time. The network consisted of a 2D generation module where oblique slices were synthesized from two orthogonal MRI slices and a 3D refinement module to generate final high quality volumetric images. A physics module that utilized prior knowledge of imaging physics was used to bridge 2D and 3D domains. The proposed method was evaluated on four patients with intrahepatic carcinoma. Each patient received two 4D MRI examinations that were more than 1 month apart, where one examination was used for network training and another for evaluation. Volumetric images of 21 motion phases were reconstructed via retrospective binning and served as the ground truth. Orthogonal Cine MRI were sampled from volumetric MRI and served as network inputs. To evaluate tracking accuracy using the network predicted volumetric MRI, a reference MRI with gross tumor volume (GTV) contours was deformed to match each of the 21 motion phases. The deformed GTV centroid positions and contours were compared between ground truth and network prediction.Across all the 4 patients evaluated, the averaged distance between predicted and ground truth GTV centroids was less than 1 mm in all 3 directions (anterior-posterior, inferior-superior and lateral). The averaged 95-percentile Hausdorff distances between the predicted and ground truth GTV contours was 4.2 mm, which was similar to the cross-plane imaging resolution (4 mm).Physics-aware deep neural network enabled real time volumetric MRI with sufficient accuracy to support 3D motion tracking. The proposed technique is robust to longitudinal patient position and configuration changes, and has the potential of reducing treatment margins and improving treatment delivery precision on a MR-Linac system.

Automated approach for segmenting gross tumor volumes for lung cancer stereotactic body radiation therapy using CT-based dense V-networks.

The aim of this study was to develop an automated segmentation approach for small gross tumor volumes (GTVs) in 3D planning computed tomography (CT) images using dense V-networks (DVNs) that offer more advantages in segmenting smaller structures than conventional V-networks. Regions of interest (ROI) with dimensions of 50 × 50 × 6–72 pixels in the planning CT images were cropped based on the GTV centroids when applying stereotactic body radiotherapy (SBRT) to patients. Segmentation accuracy of GTV contours for 192 lung cancer patients [with the following tumor types: 118 solid, 53 part-solid types and 21 pure ground-glass opacity (pure GGO)], who underwent SBRT, were evaluated based on a 10-fold cross-validation test using Dice’s similarity coefficient (DSC) and Hausdorff distance (HD). For each case, 11 segmented GTVs consisting of three single outputs, four logical AND outputs, and four logical OR outputs from combinations of two or three outputs from DVNs were obtained by three runs with different initial weights. The AND output (combination of three outputs) achieved the highest values of average 3D-DSC (0.832 ± 0.074) and HD (4.57 ± 2.44 mm). The average 3D DSCs from the AND output for solid, part-solid and pure GGO types were 0.838 ± 0.074, 0.822 ± 0.078 and 0.819 ± 0.059, respectively. This study suggests that the proposed approach could be useful in segmenting GTVs for planning lung cancer SBRT.

MR-guided Gated Stereotactic Radiation Therapy Delivery for Lung, Adrenal, and Pancreatic Tumors: A Geometric Analysis

We implemented magnetic resonance-guided breath-hold stereotactic body radiation therapy in combination with visual feedback using the MRIdian system. Both accuracy of gated delivery and reproducibility of tumor positions were studied. Tumor tracking is realized through repeated magnetic resonance imaging in a single sagittal plane at 4 frames per second with deformable image registration. An in-room monitor allowed visualization of the tracked gross tumor volume (GTV) contour and the planning target volume (PTV) (GTV+3mm), which was the gating boundary. For each delivery, a predefined threshold-region of interest percentage (ROI%)allows a percentage of GTV area to be outside the gating boundary before a beam-hold is triggered. Accuracy of gated delivery and tumor position reproducibility during breath-holds was analyzed for 15 patients (87 fractions) with lung, adrenal, and pancreas tumors. For each fraction, we analyzed (1) reproducibility of system-tracked GTV centroid position within the PTV; (2) geometric coverage of GTV area within the PTV; (3) treatment duty cycle efficiency; (4) effects of threshold ROI% settings on treatment duty cycle efficiency and GTV area coverage; and (5) beam-off latency effect on mean GTV coverage. For lung, adrenal, and pancreatic tumors, grouped 5th to 95th percentile distributions of GTV centroid positions in the dorsoventral direction, relative to PTV-center of mass (COM), were, respectively, -3.3mm to 2.8mm, -2.5mm to 3.7mm, and -4.4mm to 2.9mm. Corresponding distributions in the craniocaudal direction were -2.6mm to 4.6mm, -4.1mm to 4.4mm, and -4.4mm to 4.5mm, respectively. Mean GTV areas encompassed during beam-on for all fractions were 94.6%, 94.3%, and 95.3% for lung, adrenal, and pancreas tumors, respectively. Mean treatment duty cycle efficiency ranged from 67% to 87% for these tumors. Use of higher threshold-ROI% resulted in increased duty cycle efficiency, at the cost of a small decrease in GTV area coverage. The beam-off latency had a marginal impact on the GTV coverage. Gated stereotactic body radiation therapy delivery during breath-hold, real-time magnetic resonance guidance resulted in at least 95% geometric GTV coverage in lung, adrenal, and pancreatic tumors.

Comparison of internal tumor volume based on different reconstruction modes of 4DCT for solitary pulmonary lesion

Objective To explore the influence of different reconstruction modes with time-weighted respiratory phases on the internal tumor volume (ITV) of solitary pulmonary lesion (SPL), and to evaluate the feasibilities of 8 and 4 equal time-weighted respiratory phases in 4DCT simulation. Methods 24 patients with SPL underwent 4D scanning. Images were reconstructed with 10, 8 and 4 equal time-weighted phases of the respiratory cycles, respectively. Gross tumor volumes (GTVs) were delineated on the three sets of reconstructed images and fused into ITVs, which were ITV10, ITV8 and ITV4 respectively. The differences of volumes, centroid of the ITVs and motions of GTV centroids in three-dimensional directions were compared. Statistical analysis was performed using the Friedman M test. Results The volumes of ITV10, ITV8 and ITV4 were (9.09±12.29) cm3, (9.10±12.47) cm3 and (8.98±12.61) cm3(P=0.001), respectively. There were no differences between the volumes of ITV10 and ITV8 after the Bonferroni correction (P=0.721), while the opposite between those of ITV10 and ITV4(P=0.002). The differences of centroid positions of ITV10, ITV8 and ITV4 in x-, y-and z-axes were all less than 1 mm ((12.22±7.71), (12.23±7.71), (12.22±7.71), Px=0.668); (43.30±29.38), (43.30±29.40), (43.31±29.39), Py=0.643; (5.66±3.67), (5.66±3.67), (5.66±3.67), Pz=0.878), similar to the motions of GTV centroids in three reconstructed modes ((0.69±0.56), (0.69±0.68), (0.79±0.51) mm, Px=0.356; (3.13±3.78), (3.13±4.05), (3.19±4.06) mm, Py=0.978; (1.18±1.31), (1.03±1.32), (1.16±1.34) mm, Pz=0.302). Conclusions There were no differences in volumes, centroid positions and motions between ITV10 and ITV8. The quantity of reconstruction images and GTV delineations according to 8 time-weighted phases were both less than conventional 10 phases. 8 time-weighted respiratory phases mode was feasible in 4DCT simulation for SPL. Key words: Tomography, X-ray computed, four-dimensional; Imaging reconstruction; Internal target volume; Lung neoplasms/radiotherapy

Dosimetric study of a respiratory gating technique based on four-dimensional computed tomography in non-small-cell lung cancer

This study sought to compare the differences in target volumes and dose distributions to the targets and organs at risk (OARs) between a four-dimensional computed tomography (4DCT)-based respiratory-gated intensity-modulated radiation therapy (IMRT) plan (PlanEOE) and a three-dimensional CT (3DCT)-based IMRT plan (Plan3D) in patients with non-small-cell lung cancer (NSCLC). For 17 patients with Stages I–III NSCLC, both 4DCT data and conventional 3DCT data were obtained. The Plan3D and PlanEOE were designed based on 3DCT data and 4DCT data, respectively. The displacements of the gross tumor volume (GTV) centroid were 0.13 ± 0.09 cm, 0.15 ± 0.1 cm, and 0.27 ± 0.27 cm in the right–left, anterior–posterior, and superior–inferior directions, respectively. The volume of the GTVEOE was 3.05 ± 5.17 cm3 larger than that of the GTV3D. The volume of the PTV3D was 72.82 ± 48.65 cm3 larger than that of the PTVEOE. There was no significant difference between the PTV3D and PTVEOE for V55.8, V60, V66 and the homogeneity index. The PTV3D had a lower target conformity index than the PTVEOE (P = 0.036). PlanEOE had a significantly lower lung V10, V20, V30, V40 and mean lung dose (MLD) than Plan3D. For the heart, PlanEOE had a significantly lower V30 and mean dose. In conclusion, 4DCT is an appropriate method for assessing the displacement of the GTV centroid in three dimensions. PlanEOE has smaller PTVs and a decreased dose and volume for the normal lung and heart, as compared with Plan3D.

Detection of interfraction displacement and volume variance during radiotherapy of primary thoracic esophageal cancer based on repeated four-dimensional CT scans

BackgroundTo investigate the interfraction displacement and volume variation of primary thoracic esophagus carcinoma with enhanced four-dimensional computed tomography (4DCT) scanning during fractionated radiotherapy.Methods4DCT data sets were acquired at the time of treatment simulation and every ten fraction for each of 32 patients throughout treatment. Scans were registered to baseline (simulation) 4DCT scans by using bony landmarks. The gross tumor volumes (GTVs) were delineated on each data set. Coordinates of the GTV centroids were acquired on each respiration phase. Distance between center of the GTV contour on the simulation scan and the centers on subsequent scans were used to assess interfraction displacement between fractions. Volumes were constructed using three approaches: The GTV delineated from the maximum intensity projection (MIP) was defined IGTVMIP, all 10 GTVs were combined to form IGTV10, GTVmean was the average of all 10 phases of each GTV.ResultsInterfraction displacement in left-right (LR), anterior-posterior (AP), superior-inferior (SI) directions and 3D vector were 0.13 ± 0.09 cm, 0.16 ± 0.12 cm, 0.34 ± 0.26 cm and 0.43 ± 0.24 cm, respectively between the tenth fraction and simulation 4DCT scan. 0.14 ± 0.09 cm, 0.19 ± 0.16 cm, 0.45 ± 0.43 cm and 0.56 ± 0.40 cm in LR, AP, SI and 3D vector respectively between the twentieth fraction and simulation 4DCT scan. Displacement in SI direction was larger than LR and AP directions during treatment. For distal esophageal cancer, increased interfraction displacements were observed in SI direction and 3D vector (P = 0.002 and P = 0.001, respectively) during radiotherapy. The volume of GTVmean, IGTVMIP, and IGTV10 decreased significantly at the twentieth fraction for middle (median: 34.01%, 33.09% and 28.71%, respectively) and distal (median: 22.76%, 25.27% and 23.96%, respectively) esophageal cancer, but for the upper third, no significant variation were observed during radiotherapy.ConclusionsInterfractional displacements in SI direction were larger than LR and AP directions. For distal location, significant changes were observed in SI direction and 3D vector during radiotherapy. For middle and distal locations, the best time to reset position should be selected at the twentieth fraction when the primary tumor target volume changed significantly, and it was preferable to guide target correction and planning modification.

TH-C-213AB-01: Spatial and Dose-Response Analysis of Fibrotic Lung Changes After Stereotactic Body Radiation Therapy

Purpose: To evaluate fibrotic lung changes in patients that underwent stereotactic body radiation therapy(SBRT) for lungcancer. Specifically, we contoured volumes visually observed to have fibrotic CT changes, performed a dose‐responseanalysis, and studied the spatial location and movement of the fibrotic volumes. Methods: Our study cohort included 51 SBRT patients and 231 follow up CT scans. The follow up CT scans were binned into intervals of 3, 6, 12, 18, 24, 30, 36, 42, and 48 months post therapy. Areas of visible fibrotic changes were contoured on follow up CTs. Deformable registration was used to align each follow up CT to the pre‐treatment CT scan.Spatial analysis was performed by determining the centroid of the gross tumor volume (GTV) and each fibrosis volume and analyzing the location and motion of the fibrosis volume relative to the GTV. A dose‐responseanalysis was performed by binning each voxel in the fibrosis volume into 1 0Gy dose‐bins and analyzing the CT numbers in each dose‐bin with respect to dose and time after treatment. Results: The distribution of fibrosis volumes was 51%, 39%, and 10% in the superior, middle, and inferior regions of the lung respectively. The average radial travel between the fibrosis volume centroid and GTV centroid was 2.6±0.2cm. The volume of the fibrosis region was greatest at 3 months (47cc) and decreased over time to 18cc at 46 months. Dose‐response data revealed increasing CT numbers as a function of dose (with no decline at higher doses) and increasing CT number as a function of time. Conclusions: The highest concentration of fibrosis volumes was located in the upper third of the lung. Future work will explore the relationship between fibrosis location, fibrosis volume, and dose. Our results show consolidation of the fibrotic volume and increase of CT number with dose and time.

A correlation study on position and volume variation of primary lung cancer during respiration by four-dimensional CT

Objective To investigate the correlation of position movement of primary tumor with interested organs and skin markers,and to investigate the correlation of volume variation of primary tumors and lungs during different respiration phases for patients with lung cancer at free breath condition scanned by four-dimensional CT (4DCT) simulation.Methods 16 patients with lung cancer were scanned at free breath condition by simulation 4DCT which connected to a respiration-monitoring system.A coordinate system was created based on image of T5 phase,gross tumor volume (GTV) and normal tissue structures of 10 phases were contoured.The three dimensional position variation of them were measured and their correlation were analyzed,and the same for the volume variation of GTV and lungs of 10 respiratory phases.Results Movement range of lung cancer in different lobe differed extinct:0.8 - 5.0 mm in upper lobe,5.7 -5.9 mm in middle lobe and 10.2 - 13.7 mm in lower lobe,respectively.Movement range of lung cancer in three dimensional direction was different:z-axis 4.3 mm ± 4.3 mm＞ y-axis 2.2 mm ± 1.0 mm ＞ x-axis 1.7 mm ± 1.5 mm ( x2 =16.22,P =0.000),respectively.There was no statistical significant correlation for movement vector of GTV and interested structures (r =-0.50 - -0.01,P =0.058 - -0.961 ),nor for volume variation of tumor and lung ( r =0.23,P =0.520 ).Conclusions Based on 4DCT,statistically significant differences of GTV centroid movement are observed at different pulmonary lobes and in three dimensional directions.So individual 4DCT measurement is necessary for definition of internal target volume margin for lung cancer. Key words: Tomography, X-ray computed, four-dimensional; Lung neoplasms; Gross tumor volume; Three-dimensional motion vector

Comparison of setup error using different reference images: a phantom and lung cancer patients study

The purpose of this study was to compare setup errors obtained with kilovoltage cone-beam computed tomography (CBCT) and 2 different kinds of reference images, free-breathing 3D localization CT images (FB-CT) and the average images of 4-D localization CT images (AVG-CT) for phantom and lung cancer patients. This study also explored the correlation between the difference of translational setup errors and the gross tumor volume (GTV) motion. A respiratory phantom and 14 patients were enrolled in this study. For phantom and each patient, 3D helical CT and 4D CT images were acquired, and AVG-CT images were generated from the 4D CT. The setup errors were determined based on the image registration between the CBCT and the 2 different reference images, respectively. The data for both translational and rotational setup errors were analyzed and compared. The GTV centroid movement as well as its correlation with the translational setup error differences was also evaluated. In the phantom study, the AVG-CT method was more accurate than the FB-CT method. For patients, the translational setup errors based on FB-CT were significantly larger than those from AVG-CT in the left-right (LR), superior-inferior (SI), and anterior-posterior (AP) directions (p < 0.05). Translational setup errors differed by >1 mm in 32.6% and >2 mm in 12.9% of CBCT scans. The rotational setup errors from FB-CT were significantly different from those from AVG-CT in the LR and AP directions (p < 0.05). The correlation coefficient of the translational setup error differences and the GTV centroid movement in the LR, SI, and AP directions was 0.515 (p = 0.060), 0.902 (p < 0.001), and 0.510 (p = 0.062), respectively. For lung cancer patients, respiration may affect the on-line target position location. AVG-CT provides different reference information than FB-CT. The difference in SI direction caused by the 2 methods increases with the GTV movement. Therefore, AVG-CT should be the prefered choice of reference images.

Investigation of the Relationship Between Gross Tumor Volume Location and Pneumonitis Rates Using a Large Clinical Database of Non-Small-Cell Lung Cancer Patients

Studies have suggested that function may vary throughout the lung, and that patients who have tumors located in the base of the lung are more susceptible to radiation pneumonitis. The purpose of our study was to investigate the relationship between gross tumor volume (GTV) location and pneumonitis rates using a large clinical database of 547 patients with non-small-cell lung cancer. The GTV centroids of all patients were mapped onto one common coordinate system, in which the boundaries of the coordinate system were defined by the extreme points of each individual patient lung. The data were qualitatively analyzed by graphing all centroids and displaying the data according to the presence of severe pneumonitis, tumor stage, and smoking status. The centroids were grouped according to superior-inferior segments, and the pneumonitis rates were analyzed. In addition, we incorporated the GTV centroid information into a Lyman-Kutcher-Burman normal tissue complication probability model and tested whether adding spatial information significantly improved the fit of the model. Of the 547 patients analyzed, 111 (20.3%) experienced severe radiation pneumonitis. The pneumonitis incidence rates were 16%, 23%, and 21% for the superior, middle, and inferior thirds of the lung, respectively. Qualitatively, the GTV centroids of nonsmokers were notably absent from the superior portion of the lung. In addition, the GTV centroids of patients who had Stage III and IV clinical staging were concentrated toward the medial edge of the lung. The comparison between the GTV centroid model and the conventional dose-volume model did not yield a statistically significant difference in model fit. Lower pneumonitis rates were noted for the superior portion of the lung; however the differences were not statistically significant. For our patient cohort, incorporating GTV centroid information did not lead to a statistically significant improvement in the fit of the pneumonitis model.

Implementation of Feedback-Guided Voluntary Breath-Hold Gating for Cone Beam CT-Based Stereotactic Body Radiotherapy

To analyze tumor position reproducibility of feedback-guided voluntary deep inspiration breath-hold (FGBH) gating for cone beam computed tomography (CBCT)-based stereotactic body radiotherapy (SBRT). Thirteen early-stage lung cancer patients eligible for SBRT with tumor motion of >1cm were evaluated for FGBH-gated treatment. Multiple FGBH CTs were acquired at simulation, and single FGBH CBCTs were also acquired prior to each treatment. Simulation CTs and treatment CBCTs were analyzed to quantify reproducibility of tumor positions during FGBH. Benefits of FGBH gating compared to treatment during free breathing, as well treatment with gating at exhalation, were examined for lung sparing, motion margins, and reproducibility of gross tumor volume (GTV) position relative to nonmoving anatomy. FGBH increased total lung volumes by 1.5 times compared to free breathing, resulting in a proportional drop in total lung volume receiving 10 Gy or more. Intra- and inter-FGBH reproducibility of GTV centroid positions at simulation were 1.0 ± 0.5 mm, 1.3 ± 1.0 mm, and 0.6 ± 0.4 mm in the anterior-posterior (AP), superior-inferior (SI), and left-right lateral (LR) directions, respectively, compared to more than 1 cm of tumor motion at free breathing. During treatment, inter-FGBH reproducibility of the GTV centroid with respect to bony anatomy was 1.2 ± 0.7 mm, 1.5 ± 0.8 mm, and 1.0 ± 0.4 mm in the AP, SI, and LR directions. In addition, the quality of CBCTs was improved due to elimination of motion artifacts, making this technique attractive for poorly visualized tumors, even with small motion. The extent of tumor motion at normal respiration does not influence the reproducibility of the tumor position under breath hold conditions. FGBH-gated SBRT with CBCT can improve the reproducibility of GTV centroids, reduce required margins, and minimize dose to normal tissues in the treatment of mobile tumors.

Anatomic and Pathologic Variability During Radiotherapy for a Hybrid Active Breath-Hold Gating Technique

To evaluate intra- and interfraction variability of tumor and lung volume and position using a hybrid active breath-hold gating technique. A total of 159 repeat normal inspiration active breath-hold CTs were acquired weekly during radiotherapy for 9 lung cancer patients (12-21 scans per patient). A physician delineated the gross tumor volume (GTV), lungs, and spinal cord on the first breath-hold CT, and contours were propagated semiautomatically. Intra- and interfraction variability of tumor and lung position and volume were evaluated. Tumor centroid and border variability were quantified. On average, intrafraction variability of lung and GTV centroid position was <2.0 mm. Interfraction population variability was 3.6-6.7 mm (systematic) and 3.1-3.9 mm (random) for the GTV centroid and 1.0-3.3 mm (systematic) and 1.5-2.6 mm (random) for the lungs. Tumor volume regressed 44.6% +/- 23.2%. Gross tumor volume border variability was patient specific and demonstrated anisotropic shape change in some subjects. Interfraction GTV positional variability was associated with tumor volume regression and contralateral lung volume (p < 0.05). Inter-breath-hold reproducibility was unaffected by time point in the treatment course (p > 0.1). Increases in free-breathing tidal volume were associated with increases in breath-hold ipsilateral lung volume (p < 0.05). The breath-hold technique was reproducible within 2 mm during each fraction. Interfraction variability of GTV position and shape was substantial because of tumor volume and breath-hold lung volume change during therapy. These results support the feasibility of a hybrid breath-hold gating technique and suggest that online image guidance would be beneficial.

Lack of Correlation Between External Fiducial Positions and Internal Tumor Positions During Breath-Hold CT

For thoracic tumors, if four-dimensional computed tomography (4DCT) is unavailable, the internal margin can be estimated by use of breath-hold (BH) CT scans acquired at end inspiration (EI) and end expiration (EE). By use of external surrogates for tumor position, BH accuracy is estimated by minimizing the difference between respiratory extrema BH and mean equivalent-phase free breathing (FB) positions. We tested the assumption that an external surrogate for BH accuracy correlates with internal tumor positional accuracy during BH CT. In 16 lung cancer patients, 4DCT images, as well as BH CT images at EI and EE, were acquired. Absolute differences between BH and mean equivalent-phase (FB) positions were calculated for both external fiducials and gross tumor volume (GTV) centroids as metrics of external and internal BH accuracy, respectively, and the results were correlated. At EI, the absolute difference between mean FB and BH fiducial displacement correlated poorly with the absolute difference between FB and BH GTV centroid positions on CT images (R(2) = 0.11). Similarly, at EE, the absolute difference between mean FB and BH fiducial displacements correlated poorly with the absolute difference between FB and BH GTV centroid positions on CT images (R(2) = 0.18). External surrogates for tumor position are not an accurate metric of BH accuracy for lung cancer patients. This implies that care should be taken when using such an approach because an incorrect internal margin could be generated.

Measurement of Interfraction Variations in Position and Size of Target Volumes in Stereotactic Body Radiotherapy for Lung Cancer

To investigate the interfraction variations in volume, motion range, and position of the gross tumor volume (GTV) in hypofractionated stereotactic body radiotherapy (SBRT) for lung cancer using four-dimensional computed tomography. Four-dimensional computed tomography scans were acquired for 8 patients once at treatment planning and twice during the SBRT period using a stereotactic body frame. The image registration was performed to correct setup errors for clinical four-dimensional computed tomography. The interfraction variations in volume, motion range, and position of GTV were computed at end-inhalation (EI) and end-exhalation (EE). The random variations in the GTV were 0.59 cm(3) at EI and 0.53 cm(3) at EE, and the systematic variations were 3.04 cm(3) at EI and 3.21 cm(3) at EE. No significant variations in GTV were found during the SBRT sessions (p = .301 at EI and p = .081 at EE). The random variations in GTV motion range for the upper lobe in the craniocaudal direction were within 1.0 mm and for the lower lobe was 3.4 mm. The interfraction variations in the GTV centroid position in the anteroposterior and craniocaudal directions were mostly larger than in the right-left direction; however, no significant displacement was observed among the sessions in any direction. For patients undergoing hypofractionated SBRT, interfraction variations in GTV, motion range, and position mainly remained small. An additional approach is needed to assess the margin size.

SU-FF-T-565: Implementation of Feedback Guided Voluntary Breath Hold Gating for CBCT Based Stereotactic Body Radiotherapy

Purpose: To implement feedback guided voluntary breathhold for CBCT based stereotactic body radiotherapy.Methods and materials: 8 patients with early stage lungtumors eligible for stereotactic body radiotherapy and with respiratory induced tumor motion of > 1 cm (determined from 4D CT) were selected for feedback guided breath hold implementation. Visual feedback was provided to assist patients in maintaining consistent breath hold levels. Multiple fast spiral CT scans were acquired, both within and between successive breath holds. A single breath hold CT was chosen as the reference. Gross tumor volume (GTV) was delineated to encompass the suspected tumor volume on all breath hold CT acquisitions. Standard margins for treatment planning were 0.8 cm (CTV) and 0.3 cm (PTV). Prior to each treatment delivery, a CBCT was acquired under a series of breathholds. Changes in volume, shape and size from reference were determined using 3D‐3D matching tools. Shifts made based on 3D‐3D matching were then verified through 2D‐2D matching of acquired orthogonal MV breath hold images.Results: Overall median breath hold duration was 20 sec (max 52 sec; min 3 sec). CBCTs required 2 – 3 breathholds and were noted to be of superior quality to those obtained during normal respiration. Intra and inter breath hold reproducibility (RMS) of GTV centroid position at the time of repeat simulation was 1.2, 2.1 and 1.0 mm (AP, SI, LR). Inter fraction breath hold reproducibility of GTV position with respect to bony anatomy was 0.5,1.7 and 0.4 mm (AP, SI, LR ‐ systematic) and 1.3, 1.9 and 1.1 mm (AP, SI, LR ‐ random). Conclusions: Respiratory feedback and image guided breath hold gating can improve the accuracy/precision of stereotactic body radiotherapy, especially for mobile tumors.

Evaluation of respiratory-induced target motion for esophageal tumors at the gastroesophageal junction

Purpose To quantify the internal motion margin requirements for radiotherapy of tumors near the gastroesophageal junction (GEJ). Methods and materials Four-dimensional computed tomography (4DCT) scans were obtained for 25 patients with primary tumors located near the GEJ. The gross tumor volume (GTV) was manually contoured on the exhale-phase image from the 4DCT image set. A deformable image registration method was used to automatically propagate the contours to other phases of the 4DCT images. To quantify target motion, we measured the displacement of the GTV centroid and the variations in the target boundary and volume. Internal margins were calculated in the lateral (RL), anterior–posterior (AP), and superior–inferior (SI) directions. Results The mean ± standard deviation peak-to-peak GTV centroid motion was 0.39 ± 0.27 cm (range, 0.04–1.09 cm) in the RL, 0.38 ± 0.23 cm (range, 0.10–0.94 cm) in the AP, and 0.87 ± 0.47 cm (range, 0.43–2.63 cm) in the SI directions, respectively. On average, the internal target volume was 72% (range, 9–172%) larger than the GTV defined on a single-phase CT image. Variations in tumor boundaries due to tissue motion and deformation suggested asymmetric margins: 1.0 cm left [toward the stomach], 0.8 cm right, 1.1 cm anterior, 0.6 cm posterior, 1.0 cm superior (toward the distal esophagus), and 1.6 cm inferior (toward the stomach). Conclusion Because tumors near the GEJ are subject to a marked but asymmetric amount of respiratory-induced intrafractional tumor motion, the use of asymmetric internal margins may be beneficial.